The present invention relates to a digital video tape recorder (herein after referred to as digital VTR) having a track form at for recording digital video and audio signals in predetermined areas on oblique track, and relates to a digital VTR in which the digital video and audio signals are input in the form of a bit stream, and the bit stream is magnetically recorded and replayed (played back).
FIG. 41 is a diagram showing a track pattern of a conventional, general consumer digital VTR. Referring to the drawing, a plurality of tracks are formed on a magnetic tape 310, in a head scanning direction inclined to the tape transport direction, and digital video and audio signals are recorded therein. Each track is divided into two areas, a video area 312 for recording a digital video signal and an audio area 314 for recording a digital audio signal.
Two methods are available for recording video and audio signals on a video tape for such a consumer digital VTR. In one of the methods, analog video and audio signals are input, and recorded, using a video and audio high-efficiency encoding means; this is called a baseband recording method. In the other method, the bit stream having been digitally transmitted is recorded; this method is called a transparent recording method.
For the system of recording ATV (advanced television) signals, now under consideration in the United States, the latter, transparent recording method is suitable. This is because the ATV signal is digitally compressed signals, and does not require a high-efficiency encoding means or a decoding means, and because there is no degradation in the picture quality due to transmission.
The transparent recording system however is associated with a problem in the picture quality in a special replay mode, such as a high-speed replay mode, a still replay mode and a slow replay mode. In particular, when a rotary head scans the tape obliquely to record a bit stream, almost no image is replay at the time of high-speed replay, if not specific measure is taken.
An improvement for the picture quality for the transparent recording system recording the ATV signal is described in an article Yanagihara, et al. "A Recording Method of ATV data on a Consumer Digital VCR", in International Workshop on HDTV, 93, Oct. 26 to 28, 1993, Ottawa, Canada. Proceedings, Vol. II. This proposal is now explained.
With one basic specification of a prototype consumer digital VTR, in the SD (standard definition) mode, when the recording rate of the digital video signal is 25 Mbps, and the field frequency is 60 Hz, two rotary heads are used for recording a digital video signal of one frame, being divided into video areas on 10 tracks. If the data rate of the ATV signal is 17 to 18 Mbps, transparent recording of the ATV signal is possible with the recording rate in this SD mode.
FIG. 42A and FIG. 42B show tracks formed in a magnetic tape using a conventional digital VTR. FIG. 42A is a diagram showing scanning traces of the rotary heads during normal replay. FIG. 42B shows scanning traces of the rotary heads during high-speed replay. In the example under consideration, the rotary heads are provided in opposition, 180.degree. spaced apart on a rotary drum, and the magnetic tape is wrapped around over 180.degree.. In the drawing, adjacent tracks on the tape 310 are scanned by two rotary heads A and B having different azimuth angles, alternately and obliquely, to record digital data. In normal replay, the transport speed of the tape 310 is identical to that during recording, so that the heads trace along the recorded tracks. During high-speed replay, the tape speed is different, so that the heads A and B traces the magnetic tape 310 crossing several tracks. The arrow in FIG. 42B indicates a scanning trace by a head A at the time of five-time high-speed feeding. The width of arrow represents the width of the region read by the head. Fractions of digital data recorded on tracks having an identical azimuth angle are replayed from regions meshed in the drawings, within five tracks on the magnetic tape 310.
The bit stream of the ATV signal is according to the standard of the MPEG2. In this bit stream according to the MPEG2, only the intra-frame or intra-field encoded data of the video signal, i.e., the data of intra encoded block (intra encoded block) alone can be decoded independently, without reference to data of other frame or field. Where the bit stream is recorded in turn on the respective tracks, the recorded data are replayed intermittently from the tracks during fast replay, and the image must be reconstructed from only the intra-encoded blocks contained in the replay data. Accordingly, the video area updated on the screen is not continuous, and only the fractions of data of intra coded block are replayed, and may be scattered over the screen. The bit stream is variable-length encoded, so that it is not ensured that all the replay data over the screen is periodically updated, and the replay data of certain parts of the video area may not be updated for a long time. As a result, this type of bit stream recording system does not provide a sufficient picture quality during fast replay in order to be accepted as a recording method for a consumer digital VTR.
FIG. 43 is a block configuration diagram showing an example of recording system in a conventional digital VTR. Referring to the drawing, reference numeral 1 denotes an input terminal for the bit stream, 2 denotes an HP data format circuit, and 3 denotes a recording format circuit. Reference numeral 4 denotes a variable-length decoder, 5 denotes a counter, 6 denotes a data extractor, 7 denotes a EOB (end of block) appending circuit, and 8 denotes an output terminal.
The video area in each track is divided into a main area for recording the bit stream of the ATV signal, and copy area for recording important part (HP data) of the bit stream which are used for reconstruction of the image in fast replay. Only the intra-encoded blocks are effective during fast replay, so that they are recorded in the copy area. To reduce the data further, the only the low-frequency components are extracted from all the intra-encoded blocks, and recorded as HP data.
The bit stream of MPEG2 is input via the input terminal 2, and led to the recording format circuit 3. The bit stream from the input terminal 1 is also input to the variable-length decoder 4, and the syntax of the bit stream of the MPEG2 is analyzed, and the intra-picture data is detected, and timing signals are generated by the counter 5, and the low-frequency components of all the blocks in the intra-picture data are extracted. Furthermore, EOBs are appended at the EOB appending circuit 7, and HP data is constructed at the HP data format circuit 2. At the recording data format circuit 3, the HP data and the bit stream to be recorded in the main area are combined into a format suitable for recording in one track, and output via the output terminal 8, and respectively recorded in the main area and the copy area.
FIG. 44 shows a recording format on the tape. The combination of an alphabetic character A, B, C, and succeeding numerals 0, 1, 2 indicate the areas where HP data are recorded. Different data Ai, Bi, Ci (i=0, 1, 2, . . . ) are recorded in each tract. An identical set of data Ai, Bi and Ci are repeatedly recorded over 17 tracks within a range indicated by RP.
FIG. 45A and FIG. 45B show an example of replay system in a conventional digital VTR. FIG. 45A schematically shows normal replay. FIG. 45B schematically shows fast replay.
Separation of data from the magnetic tape during normal replay and fast replay are performed respectively in the following ways. During normal replay, the bit stream recorded in the main areas 270 is all replayed, and the bit stream from the data separation circuit 272 are sent as the normal replay data, to an MPEG2 decoder, provided outside the replay system. The HP data from the copy area 271 are discarded. During fast replay, only the HP data from the copy area 271 are collected, and sent, as fast replay data, to the decoder. At the data separation circuit 272, the bit stream from the main areas 270 is abandoned.
A method of fast replay from a track in which a main area 270 and copy areas 271 is next described. FIG. 46A shows a scanning trace of a head. FIG. 46B shows a track regions from which the replay is possible. When the tape speed is an integer multiple of the normal replay speed, if phase-locking control is conducted by an ATF (automatic track following) method or the like for tracking by moving the head itself, the head scanning is in a predetermined phase relationship with tracks having an identical azimuth. As a result, the data replayed by the head A from the tracks recorded alternately by the heads A and B, are fixed to those from the meshed regions.
In FIG. 46B, if the signal having an output level larger than -6 dB is replayed by the heads, the data is replayed by one head from the meshed tape regions. The drawing show an example of nine-time speed replay. If replay of the signals from the meshed regions is ensured at the nine-time replay, the regions are used as copy areas, and the HP data are recorded in the copy areas, so that the reading of the HP data from these regions at this speed is possible. However, reading of these signals at different speeds is not ensured. Accordingly, a plurality of areas need to be selected from the copy areas, so that the replay signals can be read at different tape speeds.
FIG. 47 shows regions where the copy areas overlap for a plurality of different replay speeds. It shows examples of scan regions for three different tape speeds, for cases where the head is in synchronism with a track of an identical azimuth. The scan regions where the reading by the head is possible at different tape speeds overlap, at some of the regions. By selecting the regions at which the overlapping occurs as the copy areas, reading of the HP data at different tape speeds can be ensured. The drawings show the regions at which overlapping occurs at the feed-forward at four-time, nine-time, 17-time speed. These scan regions are identical to those of feed-forward at -2 time, -7 time and -15 time high speeds (i.e., rewind at 2 time, 7 time and 15 time speeds).
Even though there are overlapping regions for different tape speeds, it is not possible to determine a recording pattern so that identical regions are always traced at different speeds. This is because the number of tracks crossed by the head differ depending on the tape speed. Moreover, it is necessary for the head to be capable of starting tracing at whichever identical azimuth track. For this reason, identical HP data is repeatedly recorded over a plurality of tracks, to solve the above problem.
FIG. 48 shows examples of scanning traces of the rotary head at different tape speeds. Regions 1, 2 and 3 are selected from among the overlapping regions for five-time and nine-time speeds. If identical HP data are repeatedly recorded over 9 tracks (over 9 tracks within the range indicated by RP in FIG. 48), the HP data can be read at five-time and nine-time speeds.
FIG. 49A and FIG. 49B show scanning traces at five-time speed replay. In the illustrated example, identical HP data is repeatedly recorded over five consecutive tracks (within the region indicated by RP). As will be seen from the drawings, identical HP data is recorded over the number of tracks identical to the number of times of the tape speed (i.e., 5). In either of case 1 and case 2, either the head A or B can read HP data from corresponding azimuth track. Accordingly, providing the copy areas in each track, in a number identical to the number of times of the tape speed at the fast replay, and repeatedly recording the HP data there, the copied HP data can be read at various speeds, and in either the forward or reverse direction.
In the manner described, the special replay data is recorded in the copy areas, repeatedly, to improve the picture quality during the special replay in the transparent recording system.
FIG. 50 shows a recording format on a track in a convention digital VTR. Main area 270 and copy areas 271 are provided in one track. In a consumer digital VTR, a video area in each track has 135 sync blocks (SB), and 97 sync blocks are assigned to main areas and 32 sync blocks are assigned to copy areas. The sync blocks at the regions corresponding to the 4-, 9- and 17-time speed shown in FIG. 47 are selected for the copy areas. The data rate of the main areas is about 17.46 Mbps (97.times.75.times.8.times.10.times.30), and the data rate of the copy areas where identical data is repeated 17 times is about 338.8 kbps (32.times.75.times.8.times.10.times.30/17).
The conventional VTR described above has the following problems.
In the conventional VTR, in any of the cases of the low-speed replay of 2- to 4 time speed, and the case of a fast replay of more than 9-time speed, the data of the copy areas consisting of the predetermined number of sync blocks contained in common overlapping areas is read and used for replay. As a result, the deterioration in the picture quality which is not conspicuous in a high-speed fast replay, in which the change of the scene is quick, shows up in a lower-speed replay, in which the change is of the scene is slow.
In the conventional device, the areas where the copy areas overlap are determined without taking account of the regions where the reading is possible in slow replay or still replay. As a result, when slow or still replay is conducted in the conventional device, the reading from the copy areas is not necessary ensure. Moreover, the picture is not reconstructed from only the HP data in the copy areas, so that the pictures of slow or still replay are not obtained.
When a bit stream from the main areas is used during slow or still replay, some regions may not be scanned, or the replay output may be insufficient, so that replay data is not obtained from some regions. Thus, replay of data from all the areas is not ensured, and slow or still replay pictures of good quality cannot be obtained.
In the conventional device, where each transport packet is divided and recorded in a plurality of sync blocks on the tape, the positions at which the packet is divided and the number of sync blocks into which the packet is divided are not constant because of the image compression. That is, depending on the characteristics of the picture, the amount of data contained may vary and the length of each packet may vary. For this reason, when the transport packet is divided and recorded in many sync blocks, it is affected easily by data errors for each sync block associated with the magnetic recording and replay.
More specifically, assume that a packet of a length of 188 bytes is divided and recorded in consecutive sync blocks of a length of 77 bytes. Generally, the ratio between the length of the packets and the length of the sync block is not an integer. The number of sync blocks for each packet differs. The position at which the packet is divided also varies, and accordingly, the number of sync blocks into which the packet is divided varies between 3 and 4.
When digital data is magnetically recorded or replayed, data errors for each sync block occurs. If the data in the replayed packet contains an error, it cannot be used. A packet which is divided into four sync blocks has a higher probability of being erroneous than a packet which is divided into three sync blocks.
When data for fast replay is used, by reducing the amount of data from ordinary encoded data, no control is made to maintain that the data of the image blocks is recorded at a predetermined number of sync blocks. Accordingly, when data of frame picture for high-speed replay is recorded in a plurality of sync blocks on a magnetic tape, the encoded data of the image blocks is divided at the boundaries between the sync blocks. As a result, the blocks recorded being divided is easily affected by the data errors for each sync block, associated with the magnetic recording and replay.
When image block data of a 50 byte length is recorded, it may be recorded within a single sync block, or it may be divided into two sync blocks. In comparison with the case where recording is in one sync block only, if the recording is into two sync blocks, the effect of errors for each sync block associated with recording and replay is twice.
Moreover, the positions at which the fast replay data is recorded are determined on the basis of the head scanning traces at a specific fast replay speed. As a result, fast replay is not possible at speeds other than the specific fast replay speed.
Furthermore, the copy areas where the fast replay data is recorded are disposed on the tracks such that reading from them can be made correctly. However, slow replay is not taken account of, so that it is not sure whether data is read correctly. Thus, the conventional device does not have any assurance with regard to the picture quality of slow replay.
Moreover, when still replay is selected, the replay data is not read, and no still picture is correctly displayed.
Furthermore, with regard to the speed of the fast replay in the conventional device, even where identical copy data is recorded over 17 tracks, odd-number multiple-speeds which can be selected are limited to +17-time speed, +13-time speed, +9-time speed, +5-time speed, -15-time speed, -11-time speed, -7-time speed, and -3-time speed.
In order to check all the intra-picture data, the headers of the ATV bit steams must be analyzed for each macro block.